Jahan et al. BMC Genomics (2019) 20:149 https://doi.org/10.1186/s12864-019-5524-5

RESEARCHARTICLE Open Access The NAC family transcription factor GmNAC42–1 regulates biosynthesis of the anticancer and neuroprotective in Md Asraful Jahan1, Brianna Harris2, Matthew Lowery3, Katie Coburn2, Aniello M. Infante4, Ryan J. Percifield2, Amanda G. Ammer5 and Nik Kovinich1*

Abstract Background: Glyceollins are -derived pathogen-inducible defense metabolites (phytoalexins) from soybean (Glycine max L. Merr) that have important roles in providing defense against pathogens. They also have impressive anticancer and neuroprotective activities in mammals. Despite their potential usefulness as therapeutics, glyceollins are not economical to synthesize and are biosynthesized only transiently and in low amounts in response to specific stresses. Engineering the regulation of biosynthesis may be a promising approach to enhance their bioproduction, yet the transcription factors (TFs) that regulate their biosynthesis have remained elusive. To address this, we first aimed to identify novel abiotic stresses that enhance or suppress the elicitation of glyceollins and then used a comparative transcriptomics approach to search for TF gene candidates that may positively regulate glyceollin biosynthesis. Results: Acidity stress (pH 3.0 medium) and dehydration exerted prolonged (week-long) inductive or suppressive effects on glyceollin biosynthesis, respectively. RNA-seq found that all known biosynthetic genes were oppositely regulated by acidity stress and dehydration, but known isoflavonoid TFs were not. Systemic acquired resistance (SAR) genes were highly enriched in the geneset. We chose to functionally characterize the NAC (NAM/ATAF1/2/CUC2)-family TF GmNAC42–1 that was annotated as an SAR gene and a homolog of the Arabidopsis thaliana (Arabidopsis) indole alkaloid phytoalexin regulator ANAC042. Overexpressing and silencing GmNAC42–1 in elicited soybean hairy roots dramatically enhanced and suppressed the amounts of glyceollin metabolites and biosynthesis gene mRNAs, respectively. Yet, overexpressing GmNAC42–1 in non-elicited hairy roots failed to stimulate the expressions of all biosynthesis genes. Thus, GmNAC42–1 was necessary but not sufficient to activate all biosynthesis genes on its own, suggesting an important role in the glyceollin gene regulatory network (GRN). The GmNAC42–1 protein directly bound the promoters of biosynthesis genes IFS2 and G4DT in the yeast one-hybrid (Y1H) system. Conclusions: Acidity stress is a novel elicitor and dehydration is a suppressor of glyceollin biosynthesis. The TF gene GmNAC42–1 is an essential positive regulator of glyceollin biosynthesis. Overexpressing GmNAC42–1 in hairy roots can be used to increase glyceollin yields > 10-fold upon elicitation. Thus, manipulating the expressions of glyceollin TFs is an effective strategy for enhancing the bioproduction of glyceollins in soybean. Keywords: Phytoalexin, Transcription factor, NAC, , Glyceollins

* Correspondence: [email protected] 1Division of Plant and Soil Sciences, West Virginia University, Morgantown, West Virginia 26506, USA Full list of author information is available at the end of the article

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Jahan et al. BMC Genomics (2019) 20:149 Page 2 of 21

Background AgNO3 was largely independent of ethylene signaling In 1939 K.O. Mueller et al. reported that metabolites [21]. Herbicides such as acifluorfen elicit at least in that were elicited in potato upon inoculation with an in- part via the reactive oxygen species (ROS) signaling compatible race of Phytophthora infestans subsequently pathway(s). The ups1 loss-of-function mutant of Arabi- provided resistance to a compatible race [1]. Since then, dopsis defective in ROS signaling had reduced camalexin the pathogen-inducible defense metabolites that have levels in response acifluorfen [23]. ups1 also had reduced been identified from numerous plant species have col- camalexin levels in response to Pseudomonas syringae and lectively been referred to as ‘phytoalexins’. Some phyto- P. syringae pv maculicola (Psm), suggesting a shared biotic alexins have essential roles in defending agricultural and abiotic elicitation pathway. In soybean, treatments crops against major pathogens. A classic example is the with JA, ethylene, P. sojae WGE, or hydroxyl radical (a glyceollins of soybean that provide resistance to the ROS) were highly effective at priming glyceollin biosyn- oomycete Phytophthora sojae [2–4]. For decades re- thesis in cells distal to the point of treatment, whereas SA searchers have studied the genetic regulation of phyto- was not [23, 24]. alexin elicitation by pathogens. Efforts have recently In contrast to the abiotic stresses and signaling mole- focused on identifying the transcription factors (TFs) cules that have conserved roles in eliciting phytoalexins that activate phytoalexin biosynthesis, a goal that has in response to abiotic stresses, the TFs found to regulate been confounded by the myriad of plant responses that phytoalexin biosynthesis have varied widely among plant occur synchronously in response to pathogens. Phyto- species. GaWRKY1 activated gossypol biosynthesis in alexins are biosynthetically diverse among plant species cotton [8]. GaWRKY1 transcripts were induced by methyl and include the isoflavonoid-derived glyceollins from jasmonate (MeJA) and Verticillium dahlia but not by SA soybean, the phenylpropanoid stilbenes from grapevine, or H2O2. GaWRKY1 transcripts were co-expressed both the phenolic aldehyde gossypol from cotton, the terpenoid spatially and temporally with gossypol biosynthesis genes momilactones and phytocassanes from rice, and the indole and GaWRKY1 was able to directly bind the promoter of alkaloid camalexin from Arabidopsis [5–10]. Since the TFs (+)-δ-cadinene synthase (CAD1)intheY1Hsystem.An- that activate the biosynthesis of phytoalexins in different other WRKY-family TF, namely AtWRKY33, was identified plant species belong to different gene families and/or are from Arabidopsis to directly bind and activate the promoter non-homologous, for decades an important question has of the camalexin biosynthesis gene PAD3 [25]. WRKY33 remained whether phytoalexin TFs are as diverse as the transcripts were induced by the ROS-inducing herbicide biosynthetic pathways that they regulate. Yet, several ex- paraquat, SA, and necrotrophic fungal pathogens [10]. cellent reviews highlight that phytoalexins share common GaWRKY1 and AtWRKY33 were not homologous since abiotic elicitors [11–13]. This could suggest conserved the proteins they encode had more than 20 other proteins regulatory pathways and TFs among plant species despite that were more similar by reciprocal BLASTPs. the biosynthetic heterogeneity of phytoalexins. The R2R3-type MYB TF genes VvMYB14 and VvMYB15 Highly conserved abiotic elicitors of phytoalexins in- from grapevine were co-induced with stilbene biosynthesis clude heavy metals, herbicides, and UV irradiation. genes in response to UV irradiation, wounding, and the UV elicits stilbene phytoalexins in grapevine, Cissus pathogen Plasmopara viticola [26]. The proteins directly Antarctica,andCannabis sativa [14], the bound the promoter of STILBENE SYNTHASE (STS)in and diterpenoid phytoalexins in rice [15, 16], camalexin in transient gene reporter assays using grapevine suspension Arabidopsis [17], and glyceollins in soybean [18]. In rice, cells and induced the accumulation of stilbenes when over- loss-of-function mutants of the JA biosynthesis gene allene expressed in grapevine hairy roots [26]. Homologs of oxide cyclase (aos) or jasmonic acid-amido synthetase VvMYB14 and VvMYB15 in Arabidopsis did not regulate (osjar1–2) resulted in an almost complete loss of sakurane- camalexin biosynthesis but rather cold tolerance and tin elicitation in response to UV [19]. Yet, the diterpenoid defense-induced lignification, respectively [27, 28]. Double phytoalexins of rice were not affected in JA biosynthesis and triple mutants of the Arabidopsis R2R3 MYBs mutants. Copper chloride (CuCl2) elicitation of sakuranetin, AtMYB34, AtMYB51,andAtMYB122 had reduced cama- momilactone, and diterpenoid phytoalexins in rice was lexin levels upon elicitation with UV, AgNO3, and a PAMP dramatically reduced by JA biosynthesis inhibitors [20]. The isolated from Pythium aphanidermatum (PaNie)[29]. heavy metal silver nitrate (AgNO3) elicited glyceollin However, these three MYBs were unable to bind camalexin accumulation in soybean by reducing its degradation and biosynthesis gene promoters and feeding the triple mutant by enhancing the hydrolysis of -glycoside conju- plant with a biosynthetic intermediate restored camalexin gates that compete with glyceollins for the common biosyn- accumulation, suggesting that AtMYB34, AtMYB51,and thetic intermediate [21]. AgNO3 was shown to AtMYB122 did not regulate camalexin biosynthesis directly antagonize many plant development processes by inhibiting but rather an upstream process in the elicitation pathway ethylene perception [22]. Yet, glyceollin elicitation by [29]. The constitutive overexpression of the sorghum R2R3 Jahan et al. BMC Genomics (2019) 20:149 Page 3 of 21

MYB gene yellow seed (y1) in maize resulted in the ectopic Plant materials and growth conditions accumulation of 3-deoxyanthocyanidins in vegetative tis- Soybean seeds were obtained from the USDA-GRIN soy- sues only upon challenge with Colletotrichum graminicola bean germplasm collection and from Elroy Cober (Agri- [5]. VvMYB15 and VvMYB14 were not homologs of y1 culture and Agri-Food Canada). Harosoy 63 seeds (16 since reciprocal BLASTp’s revealed 5–20 proteins that were per batch) were sterilized in 30 mL of 70% ethanol, 0.2% more similar. triton X (v/v) for 5 min on a mixer wheel, rinsed thrice RNAi silencing of the bHLH-family TF gene OsMYC2 with sterile water, and imbibed overnight. The imbibate from rice almost completely eliminated the elicitation of was then discarded to remove growth inhibitors and sakuranetin in response to JA treatment [6]. OsMYC2 seeds were transferred to water soaked sterile vermicu- directly activated the promoter of a sakuranetin biosyn- lite (250 mL in volume) in 500 mL beakers. The beaker thesis gene by transient transactivation assays in rice tops were covered with ring-shaped sterile cheese cloth leaves [6]. Transcripts of another bHLH TF gene from and covered with plastic wrap to ensure aseptic growth. rice, namely OsDPF, were inducible in rice leaves by UV, The cheese cloth permitted passage of air between plas- CuCl2 and blast infection [9]. OsDPF directly activated tic wrap and the beaker top and the ring shape permit- the promoters of phytocassane and momilactone biosyn- ted the passage of light from above the beaker. Seedlings thesis genes by transient transactivation assays in rice were grown at 22 °C under a 16 h photoperiod using − − leaves. Overexpressing OsDPF resulted in increased ex- cool white T5 fluorescent lights (500 μEm2 s 1). At pression of all diterpenoid biosynthetic genes and the ac- the first trifoliate leaf stage (~ 8 day old), seedling roots cumulation of momilactones and phytocassanes, whereas were gently rinsed with sterile water to remove vermicu- decreased levels were observed in RNAi knock-down lite and were transferred to stress treatments. lines. Two homologous JA-inducible bHLHs, TSAR1 and TSAR2, were identified to directly activate triter- Stress treatments pene saponin biosynthesis genes in Medicago trunca- For all stress treatments, the roots of five seedlings were tula [9]. TSAR1 and TSAR2 were not among the top wrapped together in a germination paper (Sartorius AG, 20 most similar proteins compared to OsDPF or Göttingen, Germany) saturated with half-strength Mura- OsMYC2, and OsDPF was only the 10th most similar shige and Skoog (MS) medium (pH 5.8) containing vita- to OsMYC2. mins and 1% (w/v) sucrose unless indicated otherwise. The ANAC-typeTFgene,AtANAC042, was identified from wrapped seedlings were transferred to a 100 mL beaker Arabidopsis by T-DNA insertion mutagenesis to have re- containing 50 mL of the above medium for the control, duced levels of camalexin biosynthesis gene expressions cold, heat, wounding and UV-C treatments. Each of the and metabolites when elicited with the ROS-inducing 100 mL beakers were then placed inside a sterile 500 mL herbicide acifluorofen, bacterial flagellin, or A. brassicicola beaker and the 500 mL beaker tops were again covered with [7]. Bacterial flagellin stimulated the accumulation of AtA- a ring-shape cut of sterile cheese cloth overlaid with plastic NAC042 transcripts at the elongation zone of the root wrap. The volume of the medium in the basin of the 100 (the site of camalexin biosynthesis), and the induction was mL beaker was maintained daily for all treatments, with the abolished in the presence of either MeJA, a general kinase exception of the dehydration treatment. For dehydration, inhibitor (K252a), or a Ca2+-chelator (BAPTA). the medium-saturated germination paper was allowed to Collectively, these studies have demonstrated that phyto- dry gradually in the 100 mL beaker containing no medium. alexin biosynthetic pathways are regulated by disparate, All seedlings were grown under the temperature and light- non-homologous TFs in different plant species, raising the ing conditions listed above unless otherwise indicated. For question of whether any TF has a conserved role in regulat- heat and cold treatments, the 500 mL beakers were trans- ing the biosynthesis of phytoalexins in plants. Here, we ferred to 37 and 15 °C, respectively. For high carbon stress, used a comparative transcriptomics approach on soybean the growth medium in the 100 mL beaker was replaced that was exposed to novel abiotic stresses and identified a with 3% sucrose in water. For flooding, control medium conserved phytoalexin regulator. was maintained up to the level of the hypocotyl-root junction throughout the 9 d treatment. For phosphate deprivation (−P), half-strength MS medium (pH 5.8) Materials and methods that lacked phosphate was used (Caisson Labs, Smithfield, Chemicals UT). For UV-C treatment, seedlings in beakers were ex- (−)- was from Dr. Paul Erhardt (University of posed to a 30 W g30 t8 germicidal light (Philips, NV) every Toledo). Soybean isoflavonoid standards were purified day for 1 h. For acidity stress, seedlings were transferred and characterized according to [21]. Isoflavone standards half-strength MS medium pH 3.0 (acidified with HCl). were from Extrasynthese (France). Solvents were LC-MS After 9 d of treatment (unless indicated otherwise), grade (Fisher). the five seedlings per treatment were unwrapped and Jahan et al. BMC Genomics (2019) 20:149 Page 4 of 21

separated, flash-frozen in liquid nitrogen, lyophilized to quality of each RNA sample was determined using an dryness, and individually ground to a fine powder and RNA Nano 6000 Chip and an Agilent 2100 Bioanalyzer stored at − 80 °C for metabolite and RNA extractions. (Santa Clara, CA). RNA samples with an Integrity Number The stored tissue powder was lyophilized again for 1 h (RIN) greater than 8.0 were used to prepare the libraries. prior to weighing. Following quantification of RNA using a Qubit fluorometer, For hairy root experiments, only secondary roots that libraries were constructed from 750 ng using the mRNA grew to 3–6 cm on selection media were considered trans- stranded library prep kit (KAPA Biosystems) as per manu- genic and were used for WGE treatments. Roots were cut facturer’s protocol with nine cycles of PCR. The completed into 1-cm pieces then overlaid with sterile water (mock) cDNA libraries were quantified using a Qubit and pooled or wall glucan elicitor (WGE) that was extracted from P. in equimolar ratios prior to sequencing at the Marshall sojae according to [21]. For RNA extraction, 100 mg of University Genomics Core. The 100 bp paired-end reads fresh tissue was harvested on ice and freeze dried prior to were generated using a HiSeq1500 system (Illumina). Eight storage at − 80 °C. For metabolite analyses, fresh hairy root libraries were sequenced per lane in high-output mode. tissues (~ 100 mg) were extracted immediately upon har- Data filtering was carried out to eliminate adapter se- vesting without lyophilization. quences and/or low-quality reads. The quality of raw reads was determined using FastQC software (http://www.bioin Isoflavonoid analysis formatics.babraham.ac.uk/projects/fastqc/) and clean reads For analysis of seedlings, lyophilized tissue powder (12 mg) were then mapped/aligned to Glycine max reference gen- − was extracted with 80% ethanol (10 μLmg 1 dry tissue) and ome (Gmax_275_V2.0.fa, https://phytozome.jgi.doe.gov/pz/ isoflavonoid identifications were done by UPLC-PDA-MSn portal.html)usingSTARRNA-seqaligner[30] with default as indicated in [21]. Four seedlings per treatment were mode based on the current gene annotation. Only the individually extracted for metabolite analysis. Metabol- paired mapped reads were considered for further analyses. ite analyses of pH 3.0 medium, dehydration stress, and Reads were quantified using using featureCounts [31]. Dif- control treatments were confirmed by three independ- ferentially expressed genes (DEGs) were identified using a ent experiments. Negative Binomial Distribution in DESeq2 [32]. Multiple − Hairy roots were extracted with 80% ethanol (1 μLmg 1 hypothesis correction was conducted with Benjamini Hoch- fresh weight, FW) as described [21]. For all hairy root ex- berg procedure to get an adjusted P value at 0.05 which de- periments, five biological replicates were analyzed per crease the false discovery rate (FDR). Principle component treatment. Two independent transformation experiments analysis, heatmap and clustering of the samples were done were analyzed per DNA construct. Absolute amounts of to check the robustness of the analysis. For the identifica- isoflavonoids were determined by comparison of the tion of gene homologs, genes were considered to be hom- UPLC-PDA peak areas to a concentration curve of puri- ologous if their predicted protein sequences were the best fied or authentic standards as described in [21]. matches in reciprocal BLASTPs.

RNA extraction and qRT-PCR Cloning Total RNA was isolated from lyophilized tissue powder The GmNAC42–1 ORF was PCR amplified from the cDNA using the Spectrum Plant Total RNA Kit (Sigma-Aldrich, of Harosoy63 seedlings treated with pH 3.0 medium (9 dat) St. Louis, MO, USA) as described [21]. Total RNA (500 ng) by the attB Adapter PCR protocol (Invitrogen, Carlsbad, was treated with DNase I (Amplification grade, Invitrogen, CA) using Phusion polymerase (Thermo Fisher Scientific) Carlsbad, CA, USA) to remove genomic DNA and cDNA and primers (Additional file 1: Table S1). The amplicon was was synthesized using SuperScript II Reverse Transcriptase cloned into the donor vector pDONR221 using BP Clonase (Invitrogen). cDNA templates were diluted 4-fold with II (Invitrogen, Carlsbad, CA) and following sequencing was water and qRT-PCR was conducted as described [21]. All LR recombined downstream of GFP in the pGWB6 vector qRT-PCR experiments included four biological replicates to assay subcellular localization and downstream of the and two technical replicates. Primers used in this study are GAL4 activation domain in the pDEST-GADT7 vector listed in Additional file 1:TableS1. for Y1H. For silencing, a 227-bp region of exon 2 of GmNAC42–1 was amplified from cDNA and BP cloned RNA-seq into pDONR221, which after sequencing was LR sub- Total RNA was extracted from the powder of individual cloned into the RNAi vector pANDA35HK. Hairpin in- seedlings as described above. Three individual seedlings tegrations were confirmed by sequencing. per stress treatment and their respective controls were used to make a total of 12 libraries for RNA-seq analysis. Soybean hairy roots RNA samples were sent to the Genomics Core Facility Transgenic soybean hairy roots were produced according of West Virginia University for library preparation. The to [33] with some modifications. Relatively large Williams Jahan et al. BMC Genomics (2019) 20:149 Page 5 of 21

82 soybean seeds without cracks were surface sterilized Yeast one-hybrid with 70% isopropyl alcohol (v/v) for 30 s and 10% com- G4DT and IFS2 promoter regions 1 and 2 flanked by mercial bleach (6.0% (v/v) sodium hypochlorite) for 5 min attL4 and attR1 recombination sites (Additional file 2: with gentle agitation, then rinsed three times in sterile Table S2) were synthesized by Genscript (Piscataway, MilliQ-filtered water (EMD Millipore, MA). Seeds were NJ) and recombined into the destination vector pMW#2 transferred to germination paper saturated with germin- (Addgene, Cambridge, MA) using LR clonase (Invitrogen, ation and co-cultivation (GC) medium (half-strength MS Carlsbad, CA). Clones were selected by colony PCR then salts (Caisson Labs, UT), 1% sucrose, pH 5.8, and MS vita- sequenced. Constructs were linearized by digestion with mins) in a sterile Petri dish and germinated for 3 d in the AflII (Thermo Scientific, Waltham, MA) prior to trans- dark, then transferred to cool white T5 fluorescent lights formation into yeast strain YM4271 (MATa, ura3–52, − (100 μEs 1 m2) at 24 °C, a condition that was used for all his3–200, lys2–801, ade2–101, ade5, trp1–901, leu2–3, 112, subsequent soybean transformation steps. tyr1–501, gal4D, gal80D, ade5::hisG)andwereselectedby Following pre-culture on LB-agar plates containing growth in dropout medium lacking histidine (SD-His). Bait − 50 mg L 1 kanamycin and hygromycin, Agrobacterium strains containing genomic integrations were confirmed by rhizogenes strain K599 containing the empty vector or colony PCR using a pair of promoter- and genome-specific construct DNA were resuspended to an OD600 of primers (Additional file 1: Table S1). Bait strains were 0.5–0.8 in phosphate buffer (0.01 M Na2HPO4,0.15M transformed with pDEST-GADT7 (ABRC, Columbus, NaCl, pH 7.5) containing 100 μM acetosyringone. Cotyle- OH) or with GmNAC42–1-pDEST-GADT7 and trans- dons were gently twisted off of 6–7 d old seedlings. The ap- formants were selected on media lacking histidine and ical meristem and hypocotyl was excised and several 1 leucine (SD-His-Leu) then confirmed by colony PCR. mm-deep cuts were made across the adaxial surface of the Autoactivation was tested for and positive DNA-protein cotyledon with a scalpel previously dipped in the Agrobac- interactions were determined by growth in SD-His-Leu terium solution. Twenty-four to 36 cotyledons were inocu- medium containing increasing concentrations (5, 10, 20, lated per DNA. Cotyledons were placed adaxial-side-down 40 and 60 mM) of 3-amino-1,2,4-triazole (3-AT; Fisher on germination paper saturated with GC medium con- Scientific, Hampton, NH), as previously described [34]. taining 100 μM acetosyringone and co-cultivated for 3 d Three biological replicates are shown, results were con- − at 22 °C under low light (65 μEs 1 m2) on a 16 h photo- firmed by two independent experiments. period. Cotyledons were then cultured adaxial-side-up on hairy root growth (HRG) medium (half strength MS salts, Results − 3% sucrose (w/v) (pH 5.8) with gelzan (2.4 g L 1; Novel abiotic stresses that regulate glyceollin biosynthesis − Sigma-Aldrich, MO), MS vitamins (2.5 mL L 1) and time- To gain insight into how abiotic stresses regulate glyceol- − ntin (500 mg L 1). Fourteen to 21 d later, transgenic pri- lin biosynthesis in soybean we first searched for a control mary roots with 2–3 cm secondary roots were transferred growth condition that would allow us to measure the in- − to and selected on HRG containing 50 mg L 1 kanamycin ductive and suppressive effects of abiotic stress treatments and hygromycin. Only secondary roots that grew to 3–6 on glyceollin biosynthesis. We grew soybean seedlings − − cm were considered transgenic and were used for treat- under two light intensities, 10 and 500 μmol m 2 s 1, ments. All hairy root experiments were conducted two which we refer to here as low and high light, respectively. times independently, representative results are shown. We also compared seedlings grown on soil to those grown in liquid half-strength Murashige and Skoog (MS) Subcellular localization medium that can be readily manipulated to provide nutri- Soybean hairy roots transformed with nGFP-pGWB6 or ent and chemical stresses (see Methods). In addition to nGFP-NAC42–1-pGWB6 were harvested and stained glyceollins, we also measured the levels of two key biosyn- with propidium iodide according to the manufacturer’s thetic intermediates, two additional phytoalexins that have instructions (Sigma-Aldrich, St. Louis, MO, USA). potent anti-pathogenic and/or medicinal activities, and Three-to-four roots per genotype per two independent two constitutively biosynthesized isoflavone-glycoside transformation events were analyzed and a representa- conjugates known to compete with glyceollins for biosyn- tive result is shown. Confocal images were acquired thetic intermediates. Specifically, we measured the levels using a Nikon A1R Si confocal laser with N-SIM-E, a of glyceollin I, glyceollin II, glyceollin III, and phaseol that TiE inverted research microscope, and NIS Elements are biosynthesized from the intermediate daidzein, and software. Imaging was performed using an Apo oil 60× βprenyl that is biosynthesized from genistein objective, plus 1.5× optical zoom, and 6× digital zoom. (Fig. 1). We also measured the levels of an unknown me- Excitation and emission spectra were 488 nm and 500– tabolite that exhibited UV absorbance properties similar 550 nm for GFP and 488 nm and 570–620 nm for propi- to isoflavonoids but did not represent any of the 57 (iso)- dium iodide, respectively. flavonoid standards that we have in our library. Jahan et al. BMC Genomics (2019) 20:149 Page 6 of 21

Intermediates phenylalanine

CHS CHR Constitutive CHI CHI Constitutive IFS IFS

: R1= -Glc, R2= -H 6’’-O-malonyldaidzin: R1= -Glc-Mal, R2= -H 6’’-O-malonylgenistin genistein daidzein 6’’-O-malonylononin: R1= -Glc-Mal, R2= -CH3 Phytoalexins G2’DT I2’H

3,9-dihydroxypterocarpan

G4DT G2DT GLS GLS prenyl genistein phaseol

or

glyceollin I glyceollin II glyceollin III Fig. 1 Soybean isoflavonoid biosynthetic pathway. Inducible phytoalexins (highlighted in blue) and constitutively biosynthesized isoflavone conjugates (highlighted in black) are derived from the isoflavone intermediates daidzein or genistein. CHS chalcone synthase, CHR chalcone reductase, CHI chalcone isomerase, IFS isoflavone synthase, G2’DT genistein 2′-dimethylallyl transferase, I2’H isoflavone 2′-hydroxylase, G4DT glycinol 4-dimethylallyl transferase, G2DT glycinol 2-dimethylallyl transferase, GLS glyceollin synthase

The MS medium high light condition was the only (Fig. 2d). Overall, pH 3.0 medium had the greatest in- condition that elicited measurable amounts of all phyto- crease in glyceollin amounts, with glyceollin III becom- alexins (Fig. 2a). The MS low light condition had greater ing 25% of the total measured isoflavonoid content. amounts of glyceollins I and II but lacked phaseol and Bprenyl-genistein, and thus may not be suitable for evalu- Acidity stress enhances and dehydration suppresses ating the specificity of the effects of abiotic stresses on glyceollin biosynthesis the glyceollin pathway. Glyceollins were absent or in Pathogens generally elicit maximum glyceollin biosyn- trace amounts in seedlings grown on soil, either under thesis within 24–48 h of inoculation, then the levels rap- the high or low light conditions. Based on these results, idly decline [4, 35]. To understand the dynamics of the we selected the MS medium high light as the control regulation of glyceollin biosynthesis by pH 3.0 medium condition to evaluate the effects of abiotic stresses on and dehydration, we measured metabolite levels at regu- glyceollin biosynthesis. lar intervals up to 9 dat. Seedlings were transferred to eight abiotic stress con- Following the transfer of seedlings to the control con- ditions and the amounts of total phytoalexins were en- dition, we observed a gradual accumulation of glyceol- hanced significantly by pH 3.0 medium, UV-C, and lins and phaseol peaking at 6 dat (Fig. 3a). In contrast, dehydration compared to the control (ANOVA, Tukey βprenyl-genistein rapidly decreased up to 3 dat then post hoc test, P < 0.01) (Fig. 2b). pH 3.0 medium stimu- remained constant thereafter. Two elicitation patterns lated the greatest increase, having 22.7-fold greater distinguished pH 3.0 medium from the control. Glyceol- amounts of total phytoalexins compared to the control lin III and phaseol exhibited sharp increases from 6 dat and significantly greater amounts compared to all other to 9 dat, whereas glyceollins I and II exhibited delayed treatments. UPLC-PDA chromatograms revealed major and prolonged accumulation (Fig. 3a). Elicitation of increases in the levels of glyceollins for pH 3.0 medium, these daidzein-derived phytoalexins was accompanied by and major reductions in the amounts of 6-O-malonyl- decreases in daidzein and its glycosyl-conjugates, namely daidzin for dehydration and pH 3.0 medium that were daidzin and 6-O-malonyldaidzin. Genistein and derived not observe for the UV-C treatment (Fig. 2c). pH 3.0 isoflavonoids were not increased by pH 3.0 medium. In medium and dehydration predominantly caused in- sharp contrast, dehydration caused a sustained suppres- creases in the amounts of glyceollin III and glyceollin II sion of all daidzein-derived isoflavonoids over the 9 d Jahan htaeisrpeettesmo lcoln -I hso,and phaseol, I-II, glyceollins mean. of of sum error the standard represent represent phytoalexins bars Error seedlings. 63 Harosoy 2 Fig. rprin ftemtbltslbldi ae ntertsu ocnrto ( concentration tissue their on based C in labeled metabolites the of Proportions test, hoc post Tukey ANOVA, ab Amount ( µg g-1 FW) 100 150 200 250 300 350 400 450 50 ta.BCGenomics BMC al. et 0 htaei n sfaoodcneti epnet boi tess fet flgtadgot eimo htaei lctto in elicitation phytoalexin on medium growth and light of Effects a stresses. abiotic to response in content isoflavonoid and Phytoalexin

cd -1

Amount (µg g FW)12

Absorbance 283 nm (mAU ) 0 3 6 9 100 150 200 250 300 350 100 150 200 250 300 350 100 150 200 250 300 350 100 150 200 250 300 350 50 50 50 50 0 0 0 0 1012141618 8 6 4 2 0 Dehydration UV-C Control pH 3.0 pH (2019)20:149

P unknown 1 unknown 1 unknown 1 unknown 1 .1 ro asrpeetsadr ro fmean. of error standard represent bars Error 0.01. < Inset daidzin daidzin daidzin daidzin

-O-malonyldaidzin -O-malonyldaidzin -O-malonyldaidzin -O-malonyldaidzin

-O-malonylgenistin -O-malonylgenistin -O-malonylgenistin -O-malonylgenistin Metabolite Time (min) -O-malonylononin -O-malonylononin -O-malonylononin -O-malonylononin daidzein daidzein daidzein daidzein

genistein genistein genistein genistein glyceollin III glyceollin III glyceollin III glyceollin III glyceollin II glyceollin II glyceollin II glyceollin II glyceollin I glyceollin I glyceollin I glyceollin I β rnlgnsen ifrn etr hwsgiiatdfeecsb igefactor single by differences significant show letters Different prenyl-genistein. b fet fdfeetaitcsrs ramnspyolxneiiain Total elicitation. phytoalexin treatments stress abiotic different of Effects MS medium high light high medium MS light low medium MS high light Soil Soil low light -genistein prenyl-genistein prenyl-genistein prenyl-genistein prenyl phaseol phaseol phaseol phaseol μ Dehydration Control gg 30 c 2 1 pH 3.0 pH UV-C 16 PCPAcrmtgaso taoi xrcsa 8 nm. 283 at extracts ethanolic of chromatograms UPLC-PDA 14 − 26 0 14 1 Proportion (% (% µg Proportion g 13 DW) 1 6 5 0 -1

5 Total Phytoalexin (µg g FW) 4 0 100 150 200 250 2 3 50 2 1 0 5 3 2 2% phytoalexins 3 3 11% phytoalexins 5 18% phytoalexins 26 0 5 c -1 4 52 45% phytoalexins 7 10 1 DW) 57 4 11 0 2 0 7 5 3 c 7 3 8 25 c Abiotic stress treatment Abiotic stress c phaseol III glyceollin II glyceollin I glyceollin unknown 1 unknown genistein daidzin daidzein cc prenyl-genistein -O-malonylgenistin -O-malonyldaidzin -O-malonylononin b ae7o 21 of 7 Page b a d Jahan et al. BMC Genomics (2019) 20:149 Page 8 of 21

a 100 100 glyceollin I glyceollin II 200 glyceollin III Control pH 3.0 Dehydration

50 50 100

0 0 0 1357913579 13579 100 40 phaseol prenyl-genistein -O-malonylononin 4

50 20 FW) 1

- 2

0 0 0 135791357913579 60 60 600 daidzein daidzin -O-malonyldaidzin

40 40 400 Isoflavonoid amount (µg g (µg amountIsoflavonoid 20 20 200

0 0 0 135791357913579 20 150 150 genistein -O-malonylgenistin unknown I

100 100 10 50 50

0 0 0 1357913579 13579 b Days after treatment (dat)

9 dat ) 0.08 IFS1 a 0.20 IFS2 0.25 I2'H a 0.04 G4DT a a 0.20 0.06 0.15 0.03 0.15 0.04 0.10 0.02 UBIQUITIN3 0.10 0.02 0.05 0.05 0.01 Log2 Fold Change (Gene / (Gene / 0.00 0.00 0.00 0.00

0.12 0.35 0.25

) G4DT 6 dat IFS1 IFS2 I2'H 0.04 0.30 0.10 0.20 0.25 0.08 0.03 0.20 b 0.15 0.06 b b 0.02 UBIQUITIN3 0.15 0.10 0.04 b b 0.10 0.05 0.01

Log2 Fold Change 0.02 0.05 b (Gene / 0.00 0.00 0.00 0.00

Fig. 3 (See legend on next page.) Jahan et al. BMC Genomics (2019) 20:149 Page 9 of 21

(See figure on previous page.) Fig. 3 Time course of phytoalexin and isoflavonoid biosynthesis during acidity and dehydration stresses. a Isoflavonoid levels by UPLC-PDA over time after transfer to the control condition, pH 3.0 medium, or dehydration stress. Error bars represent standard error of mean. b Isoflavonoid biosynthesis gene expressions at 6 and 9 dat measured by qRT-PCR. aSignificantly greater and bsignificantly less than control, paired students t-test (P < 0.01). Error bars represent standard error of mean period with up to a 106.8-fold suppression of glyceollin I required for the activation of those biosynthesis genes at 6 dat (Fig. 3a). This major suppressive effect was not would also be present in our geneset. Yet, all previously observed for genistein-derived metabolites. identified isoflavonoid TF genes were not found. Those To determine whether pH 3.0 medium and dehydra- absent included TF genes identified by QTL mapping of tion stresses regulated glyceollin biosynthesis gene tran- isoflavonoid amounts, namely GmMYBJ3 (Glyma.06 scripts, we measured the expression of key biosynthetic g193600) or GmMYB29 (Glyma20g35180) [39, 40]. Also genes by quantitative reverse transcriptase-polymerase absent were TFs that activated the biosynthesis of chal- chain reaction (qRT-PCR). Specifically, we measured the cone synthase-derived isoflavonoids during seed develop- expressions of isoflavone synthase 1 (IFS1)andIFS2, iso- ment, namely GmMYB176 (Glyma.05G032200) and genes for the biosynthesis of (Fig. 1). We also GmCYP1 (Glyma.11G098700) [41, 42]. measured the expressions of isoflavone 2′-hydroxylase (I2’H) and glycinol 4-dimethylallyltransferase (G4DT), Comparative transcriptomics identifies candidate genes for the biosynthesis of all daidzein-derived phyto- transcription factors for the regulation of glyceollin alexins and glyceollin I, respectively [36, 37]. biosynthesis pH 3.0 medium upregulated all gene transcripts at 9 To better understand the pathways that were oppositely dat. The levels ranged from 4.4- to 20.7-fold greater than regulated by acidity and dehydration stresses, we ana- the control for I2’H and IFS2, respectively (Fig. 3b). By lyzed the ontologies of the 1058 oppositely regulated contrast, dehydration stress had reduced levels of all genes (Fig. 4a). Signal transduction was the most com- gene transcripts at 6 dat, ranging from 2.2- to 11.7-fold mon category of ontology (31.4% of genes, Fig. 4b). less than the control for IFS2 and I2’H, respectively. When the signal transduction category was broken down into ontologies, the greatest proportion (28.3%) Acidity and dehydration stresses oppositely regulate all were annotated as systemic acquired resistance (SAR) known glyceollin biosynthesis genes (Fig. 4c). SAR is a component of the plant immune sys- To investigate whether pH 3.0 medium and dehydration tem whereby tissues distant from a pathogen infection oppositely regulated all known glyceollin biosynthesis site become primed (sensitized) to more rapidly activate genes, we conducted RNA-seq comparing genes upregu- resistance responses the second time the plant encoun- lated by pH 3.0 medium to those downregulated by ters the pathogen. Gene ontology (GO) enrichment dehydration. analysis indicated that SAR genes were significantly − pH 3.0 medium upregulated 3242 and dehydration enriched (P <1.0 10) and included those involved of downregulated 9129 genes more than 2-fold, respectively salicylic acid (SA)-dependent and independent signaling (P < 0.05) (Additional file 3: Table S3 and Additional file 4: pathways, in addition to jasmonic acid (JA) and ethylene sig- Table S4). By comparing the two gene lists, we found that naling pathways (GO:0009627, GO:0009862, GO:0009864, 1058 genes were in common (Fig. 4a & Additional file 5: GO:0009871, and GO:0010112). The SAR genes included Table S5). All 27 known glyceollin biosynthesis genes homologs of AGD2-LIKE DEFENSE RESPONSE PROTEIN spanning from phenylalanine ammonia lyase (PAL)tothe 1 (ALD1)andFLAVIN-DEPENDENT-MONOOXYGEN- glycinol:dimethylallyl diphosphate (DMAPP) transferases ASE1 (FMO1) that were indispensable for SAR in Arabidop- G4DT and G2DT [37, 38] were upregulated by pH 3.0 sis (Table 2)[43–45]. ALD1 encodes an enzyme that medium and downregulated by dehydration, respectively synthesizes the non-protein amino acid pipecolic acid (Pip) (Table 1). Since DMAPP is derived from either the cyto- fromLysuponpathogenattack[45]. FMO1 converts Pip to solic mevalonate pathway or the plastidic methylerythritol N-hydroxypipecolic acid (NHP) [46] and is needed for Pip phosphate (MEP) pathway, we checked our lists for these to orchestrate priming of pathogen responses by SA- genes. pH 3.0 and dehydration stresses oppositely regu- dependent and independent pathways [47]. The SAR lated genes for all steps of the MEP pathway up to genes also included homologs of signaling and TF genes DMAPP formation, whereas no mevalonate genes were that had roles in regulating the elicitation of the indole differentially regulated (Table 1). alkaloid phytoalexin camalexin in Arabidopsis. PHYTO- Since our RNA-seq analyses found that pH 3.0 medium ALEXIN DEFICIENT4 (PAD4) is a lipase-like gene re- and dehydration regulated glyceollin biosynthesis at the quired for SA-dependent elicitation of camalexin in level of transcription, we hypothesized that TF genes response to microbial pathogens [48]. SIGMA FACTOR Jahan et al. BMC Genomics (2019) 20:149 Page 10 of 21

5.9 a b 1.6 6.9 c 2.0 1.6 2.0 2.7 2.6 3.7 3.3 28.3 31.4 3.7 8.6 4.3 2182 1058 8070 4.3 10.5 6.9

17.0 12.5 24.3 16.0 pH 3.0 upregulated Dehydration downregulated Signal transduction Transport Systemic acquired resistance Jasmonic acid (3242) (9129) Carbohydrate metabolism Cell differentiation Intracellular Salicylic acid Lipid metabolism Cellular metabolism Transmembrane receptor kinase Induced systemic resistance Biosynthetic process Multicellular development Resistance gene-independent Cytokinin Growth Cell death Resistance gene-dependent Others Cell growth Others Fig. 4 Comparative transcriptomics of seedlings treated with acidity stress or dehydration. a Number of genes in Harosoy 63 seedlings that were upregulated and downregulated more than 2-fold by pH 3.0 medium and dehydration, respectively (P < 0.05) by RNA-seq. b Percent of genes upregulated by pH 3.0 medium and downregulated by dehydration stress assigned to a category of gene ontology. c Breakdown of the ‘Signal transduction’ category into gene ontologies. Ontology analysis was conducted using the SoyBase Gene Model Data Mining and Analysis tool

BINDING PROTEIN 1 (SIB1)encodesaTFthatactivates (Additional file 6: Table S6). A phylogenetic analysis of the the expression of AtWRKY33, a direct regulator of predicted GmNAC42 proteins with characterized NACs camalexin biosynthesis genes [49]. However, homologs revealed that the GmNAC42s were most closely related to of AtWRKY33 (namely Glyma.02G232600 and Gly- VvNAC42_5 (Fig. 5c). VvNAC42_5 is an SA-independent ma.14G200200) were not found in our gene set nor were powdery mildew responsive gene from grapevine (Vitis vi- they significantly upregulated by pH 3.0 medium alone. nifera)[50]. Also in this cluster were proteins that posi- Among the putative soybean SAR genes were three tively regulate drought stress responses, namely SlJUB1 homologs of the NAC [no apical meristem (NAM), Arabi- and DlNAC1 [51, 52]. dopsis transcription activation factor [ATAF1/2]andcup- To probe further whether GmNAC42s may be positive shaped cotyledon (CUC2)] family gene ANAC042/AtJUB1. regulators of glyceollins, we assessed whether their gene ANAC042/AtJUB1 regulates camalexin biosynthesis in expressions were upregulated by the wall glucan elicitor Arabidopsis in response to the ROS-inducing herbicide (WGE) from P. sojae. acifluorofen, Alternaria brassicicola,andbacterialflagellin Treatment of soybean hairy roots with WGE resulted in (Flg22) [7]. maximum accumulation of glyceollins at 24 h after treat- ment (Fig. 5d). qRT-PCR found that all three GmNAC42s NAC42-type TFs are upregulated with glyceollins by were upregulated 9.6- to 14.4-fold at this time with the abiotic and biotic elicitors glyceollin biosynthesis gene G4DT (Fig. 5e). GmNAC42–1 We conducted qRT-PCR to gain insight into whether was the most highly upregulated. the NAC42-type TFs that were identified by our tran- scriptomics analysis may be involved in regulating gly- GmNAC42–1 regulates glyceollin biosynthesis in response ceollin biosynthesis. qRT-PCR confirmed that the three to Phytophthora sojae WGE GmNAC42s were upregulated by pH 3.0 medium and We chose to investigate the function of GmNAC42–1 downregulated by dehydration (Fig. 5a-b). since it is the soybean homolog of ANAC042, an indole The predicted GmNAC42 proteins were 68.5–85.8% alkaloid phytoalexin regulator from Arabidopsis, and similartoeachotherand54.3–56.7% similar to ANAC042/ since its gene expressions coincided with the elicitation JUB1 with GmNAC42–1 being the most similar of glyceollin biosynthesis. If GmNAC42–1 positively (Additional file 6: Table S6). The N-terminal halves regulates glyceollin biosynthesis, silencing its gene ex- of these proteins contained the conserved NAM do- pressions in elicited tissues should reduce the accumu- main (pfam02365) putatively involved in dimerization lation of glyceollin metabolites and biosynthesis gene and binding DNA (Additional file 7:Fig.S1).The transcripts. Conversely, overexpressing GmNAC42–1 N-terminal halves of the GmNAC42s were highly should increase the accumulation of glyceollins and their similar to ANAC042/JUB1 (76.2–83.3%), whereas the biosynthesis gene transcripts. To test, we produced soy- C-terminal halves putatively involved in protein-protein bean hairy roots harboring an RNA interference (RNAi) interactions were highly divergent (30.5–34.9% similarity) construct that encoded a hairpin dsRNA identical to a Table 1 Glyceollin biosynthesis genes upregulated by pH 3.0 medium and downregulated by dehydration Jahan Pathway Gene symbol Enzyme Wm82.a2 (Glyma 2.0) pH 3.0 medium Dehydration ta.BCGenomics BMC al. et Log2 FC P-value Log2 FC P-value Phenylpropanoid PAL phenylalanine ammonia-lyase Glyma.02G309300 1.49 2.00E-05 −2.15 1.69E-10 Glyma.03G181700 1.31 0.01535019 −2.08 0.00169183 Glyma.03G181600/ 1.10 0.00697352 −1.83 0.00121303 Glyma.10G058200 (2019)20:149 C4H cinnamic acid 4-hydroxylase Glyma.20G114200 4.44 1.63E-170 −1.78 0.00085225 4CL 4-coumarate: coenzyme A ligase Glyma.11G010500 2.17 2.39E-24 −1.08 0.00147188 CHR chalcone reductase Glyma.14G005700 2.03 3.46E-10 −1.64 0.00053532 Glyma.02G307300 3.78 6.27E-25 −3.25 1.27E-07 CHI chalcone isomerase Glyma.10G292200 1.57 3.94E-11 −4.72 4.19E-43 Glyma.20G241500 1.94 4.17E-19 −1.32 0.02583103 Isoflavonoid IFS1 Isoflavone synthase Glyma.07G202300 1.97 1.95E-13 −2.12 6.75E-15 IFS2 Isoflavone synthase Glyma.13G173500 3.23 2.16E-24 −1.11 3.40E-06 HIDH 2-hydroxy-isoflavanone dehydratase Glyma.10G250300 1.51 2.00E-09 −1.68 1.30E-05 I2’H isoflavone 2′-hydroxylase Glyma.15G156100 4.32 1.43E-81 −4.08 7.32E-18 IFR isoflavone reductase Glyma.11G070500 2.31 1.52E-29 −4.23 1.34E-11 Glyma.11G070600 2.67 1.06E-60 −3.28 4.13E-18 Glyma.01G172600 1.55 1.52E-11 −3.15 4.23E-13 Glyma.01G172700 1.80 6.88E-13 −2.96 1.54E-20 Glyma.01G211800 2.48 9.19E-11 −2.56 1.05E-06 Glyma.11G070200 2.64 1.04E-31 −1.43 9.61E-07 PTS1 pterocarpan synthase Glyma.19G151200 3.67 1.31E-20 −2.84 4.06E-06 Glyma.19G151100 3.78 1.55E-32 −1.53 0.0105246 Glyma.03G147700 3.75 1.34E-30 −1.70 0.00122376 P6αH dihydroxypterocarpan- 6α-hydroxylase Glyma.19G144700 1.71 4.51E-16 −3.79 3.21E-16 G4DT glycinol 4-dimethylallyl-transferase Glyma.10G295300 2.93 1.01E-19 −2.91 2.26E-10 G2DT glycinol 2-dimethylallyl-transferase Glyma.20G245100 2.80 5.96E-45 −3.11 3.78E-10 Methylerythritol DXS 1-deoxy-D-xylulose 5-phosphate Glyma.18G148700 2.73 2.90E-32 −4.63 4.62E-22 phosphate (MEP) synthase Glyma.08G277000 3.15 1.45E-22 −3.33 8.88E-17

Glyma.08G277100 2.49 1.07E-18 −3.01 5.41E-09 21 of 11 Page DXR 1-deoxy-D-xylulose 5-phosphate Glyma.16G089000 1.07 3.73E-10 −2.68 4.07E-07 reductoisomerase Glyma.17G089600 1.49 6.09E-20 −2.34 2.36E-08 Glyma.05G037500 1.18 1.29E-23 −1.95 1.59E-07 Table 1 Glyceollin biosynthesis genes upregulated by pH 3.0 medium and downregulated by dehydration (Continued) Jahan Pathway Gene symbol Enzyme Wm82.a2 (Glyma 2.0) pH 3.0 medium Dehydration ta.BCGenomics BMC al. et Log2 FC P-value Log2 FC P-value CMK 4-diphosphocytidyl-2-C-methyl- Glyma.20G046800 1.33 1.09E-06 −2.06 1.27E-10 D-erythritol kinase MDS 2-C-methyl-D-erythritol 2, Glyma.11G021400 1.10 4.90E-11 −1.46 6.26E-05 4-cyclodiphosphate synthase

HDS 4-hydroxy-3-methylbut-2-enyl Glyma.13G326400 1.82 9.17E-34 −1.19 1.63E-09 (2019)20:149 diphosphate synthase HDR 4-hydroxy-3-methylbut-2-enyl Glyma.11G120900 1.90 3.48E-25 −1.16 1.35E-10 diphosphate reductase Glyma.12G046000 1.22 6.42E-15 −1.09 1.61E-05 IDI2 IPP isomerase Glyma.18G242300 1.55 8.00E-07 −1.10 1.05E-06 ae1 f21 of 12 Page Jahan

Table 2 Select SAR genes upregulated by pH 3.0 medium and downregulated by dehydration Genomics BMC al. et Accession Arabidopsis Annotation Arabidopsis pH 3.0 medium Dehydration BLASTP E-value − (Wm82.a2.v1) gene symbol accession score (E < 10 6) Log2 FC P-value Log2 FC P-value Glyma.08G180600 ALD1 AGD2-LIKE DEFENSE RESPONSE PROTEIN 1 AT2G13810 3.848 0.000 −6.507 0.000 0.0E+ 00 100 Glyma.17G046600 FMO1 FLAVIN-DEPENDENT MONOOXYGENASE 1 AT1G19250 6.575 0.000 −3.981 0.000 0.0E+ 00 100

Glyma.06G156300 PAD4 PHYTOALEXIN DEFICIENT 4 AT3G52430 1.809 0.000 −1.753 0.001 0.0E+ 00 100 (2019)20:149 Glyma.03G249100 SIB1 SIGMA FACTOR BINDING PROTEIN 1 AT3G56710 4.133 0.000 −5.521 0.000 6.0E-90 100 Glyma.04G134200 2.034 0.000 −1.140 0.007 6.0E-89 100 Glyma.02G284300 ANAC042/JUB1 NAC DOMAIN CONTAINING PROTEIN AT2G43000 1.178 0.000 −2.568 0.000 0.0E+ 00 100 42/JUNGBRUNNEN 1 Glyma.18G110700 1.846 0.000 −1.527 0.003 0.0E+ 00 100 Glyma.14G030700 3.029 0.000 −1.197 0.011 0.0E+ 00 100 Glyma.13G267700 WRKY70 WRKY DNA-BINDING PROTEIN 70 AT3G56400 4.660 0.000 −3.175 0.005 0.0E+ 00 100 Glyma.13G267500 3.029 0.000 −5.123 0.000 0.0E+ 00 100 Glyma.18G213200 1.086 0.000 −4.386 0.000 0.0E+ 00 100 Glyma.04G061400 WRKY40 WRKY DNA-BINDING PROTEIN 40 AT1G80840 2.140 0.000 −2.580 0.000 1.0E-162 100 Glyma.14G103100 1.712 0.000 −5.104 0.000 0.0E+ 00 100 Glyma.11G120400 DIR1 DEFECTIVE IN INDUCED RESISTANCE 1 AT5G48485 1.327 0.000 −4.214 0.000 2.0E-68 100 Glyma.12G045500 1.220 0.002 −3.198 0.000 2.0E-67 100 Glyma.06G137000 DMR6 DOWNY MILDEW RESISTANT 6 AT5G24530 1.428 0.000 −4.998 0.000 0.0E+ 00 100 Glyma.14G048900 EFE ETHYLENE FORMING ENZYME AT1G05010 2.006 0.000 −1.333 0.017 0.0E+ 00 100 Glyma.08G128900 EFR EF-TU RECEPTOR AT5G20480 2.574 0.000 −2.045 0.007 0.0E+ 00 100 Glyma.07G103700 YLS9 NDR1/HIN1-LIKE 10 AT2G35980 3.041 0.000 −3.884 0.000 2.0E-148 100 Glyma.07G103800 1.466 0.001 −2.917 0.002 8.0E-152 100 Glyma.09G003100 RLK1 RECEPTOR-LIKE PROTEIN KINASE 1 AT5G60900 2.462 0.000 −2.888 0.000 0.0E+ 00 100 Glyma.03G090200 2.254 0.000 −1.321 0.000 0.0E+ 00 100 Glyma.15G258400 1.753 0.012 −4.378 0.000 0.0E+ 00 100 Glyma.16G168700 RLP19 RECEPTOR LIKE PROTEIN 19 AT2G15080 3.743 0.000 −5.408 0.000 0.0E+ 00 99.61 Glyma.16G169500 RLP32 RECEPTOR LIKE PROTEIN 32 AT3G05650 1.890 0.002 −2.689 0.001 0.0E+ 00 100 Glyma.16G170700 2.118 0.023 −3.478 0.001 0.0E+ 00 98.59 Glyma.16G175100 RLP33 RECEPTOR LIKE PROTEIN 33 AT3G05660 6.601 0.000 −5.803 0.000 0.0E+ 00 100 ae1 f21 of 13 Page Glyma.14G046000 3.548 0.000 −1.017 0.025 0.0E+ 00 99.76 Glyma.16G126100 RLP46 RECEPTOR LIKE PROTEIN 46 AT4G04220 2.151 0.017 −2.953 0.022 0.0E+ 00 93.3 Jahan et al. BMC Genomics (2019) 20:149 Page 14 of 21

acGmNAC42-1 GmNAC42-2 GmNAC42-3 G4DT 0.0035 0.003 0.005 0.04 a ) a a 0.003 0.0025 a 0.004 0.03 0.0025 0.002 0.002 0.003 0.0015 0.02

UBIQUIITIN3 0.0015 0.002 0.001 0.001 0.01 0.001 Log2 Fold Change 0.0005 0.0005 (Gene / 0 0 0 0 b GmNAC42-1 GmNAC42-2 GmNAC42-3 G4DT 0.007 0.003 0.09

) 0.007 0.08 0.006 0.0025 0.006 0.07 0.005 0.002 0.005 0.06 0.004 0.05 0.0015 0.004 b 0.04 UBIQUIITIN3 0.003 0.003 0.001 0.03 0.002 0.002 b 0.02 Log2 Fold Change 0.0005 0.001 b 0.001 0.01 b (Gene / 0 0 0 0

deWGE GmNAC42-1 GmNAC42-2 GmNAC42-3 G4DT ) 0.005 0.01 0.001 40 H2O a a a 0.05 a 0.004 0.008 0.0008 30 0.04 FW) 0.003 0.006 0.0006 1 - 0.03 20 UBIQUIITIN3 0.002 0.004 0.0004

(µg g 0.02

Total glyceollins Total 10 Log2 Fold Change 0.001 0.002 0.0002 0.01 (Gene / 0 0 0 0 0 0 24 48 72 96 120 144 168 192 Time after treatment (h) Fig. 5 GmNAC42s expressions are induced with glyceollins in response to abiotic and biotic elicitors. a, b Gene expressions following treatment with pH 3.0 medium at 9 dat or dehydration at 6 dat relative to their respective controls measured by qRT-PCR. aSignificantly greater and bsignificantly less than control, paired students t-test (P <0.01).c Unrooted phylogenetic tree of GmNAC42 amino acid sequences and characterized NACs. GenBank Accessions and Phytozome landmarks: GmNAC42–1 (KRH73619; Glyma.02G284300), GmNAC42–2 (KRH14512; Glyma.14G030700), GmNAC42–3 (KRG98971; Glyma.18G110700), VvNAC042_5 (XP_002283251; VIT_12s0028g00860), ANAC042 (Q9SK55; AT2G43000), SlJUB1 (XP_019069297; Solyc05g021090); AtLOV1/ ANAC035 (Q9ZVP8.2; AT2G02450), DlNAC1 (ABQ96120; N/A), ANAC032 (AAM65083; AT1G77450), GmNAC32–1 (NP_001236871.2; Glyma.06G114000), GmNAC32–2 (NP_001240958.1; Glyma.04G249000), StNTP1 (AGY49284; PGSC0003DMT400001498), StNTP2 (AGY49285; PGSC0003DMT400079997), NbNTP2 (AGY49287; N/A), SlSRN1 (NP_001304297; Solyc12g056790), GmNTP1–1 (NP_001276287; Glyma.02G222300), GmNTP1–2 (XP_006596412; Glyma.14G189300), GmNTP2–1 (XP_003556180; Glyma.20G172100), GmNTP2–2 (NP_001274375; Glyma.10G219600). The number adjacent branches indicate maximum parsimony bootstrap values for the corresponding node. The scale bar indicates the number of differences per 100 residues derived from the Muscle alignment. The phylogenetic tree was generated using MEGA v5.0 software [77]. d Amount of total glyceollins extracted from empty-vector transformed soybean hairy roots following treatment with H2O or wall glucan elicitor (WGE) from P. sojae. e Gene expressions following 24 h of WGE or water treatment of soybean hairy roots measured by qRT-PCR

227bpregionofexon2ofGmNAC42–1 and roots that resulted in 10.8-, 4.9-, and 3.0-fold increases in the overexpressed the GmNAC42–1 open reading frame amounts of glyceollin I, genistein, and glyceollin II, re- (ORF) via the constitutive cauliflower mosaic virus pro- spectively (Fig. 6e). It also caused a 1.6-to 2.7-fold reduc- moter (p35S). tion in the amounts of daidzin and 6-O-malonyldaidzin, A 2.0-fold silencing of GmNAC42–1 decreased the ac- consistent with upregulating G4DT (Fig. 6b). In the ab- cumulations of glyceollin biosynthesis gene transcripts sence of WGE treatment, the overexpression of IFS1, IFS2,andG4DT 1.8- to 2.4-fold (Fig. 6a). Off-target GmNAC42–1 alone was not sufficient to stimulate gly- silencing of GmNAC42–2 was observed but not for ceollin accumulation, reflecting its inability to upregulate GmNAC42–3. The overexpression of GmNAC42–1 upreg- all glyceollin-specific biosynthesis genes when overex- ulated IFS1, IFS2,andG4DT from 2.1- to 8.3-fold in roots pressed. However, it did result in a 2.4-fold reduction in treated with WGE or mock (H2O) (Fig. 6b-c). the amounts of 6-O-malonyldaidzin [21]. RNAi silencing of WGE-elicited roots decreased the amounts of glyceollin I, II and III 4.0-, 2.8- and 3.2-fold, GmNAC42–1 localizes to the nucleus and directly binds respectively (Fig. 6d). It also caused 2.0-fold decreases in the promoters of glyceollin biosynthesis genes the amounts of 6-O-malonyldaidzin and daidzin (Fig. 6d), To determine whether the subcellular localization of the consistent with decreased expressions of IFS2 (Fig. 6a). GmNAC42–1 protein was consistent with its putative Overexpressing GmNAC42–1 in WGE-elicited roots role as a TF, we cloned its ORF downstream of an Jahan et al. BMC Genomics (2019) 20:149 Page 15 of 21

a WGE RNAi-vector RNAi-GmNAC42-1 GmNAC42-1 IFS1 IFS2 I2'H G4DT GmNAC42-2 GmNAC42-3

) 0.0035 0.09 0.06 0.12 0.004 0.0003 0.02 0.003 0.05 0.1 0.0025 0.015 0.003 0.06 0.04 0.08 0.0002 0.002 0.06 b 0.002 UBIQUIITIN3 b 0.01 0.03 0.0015 b b 0.03 0.02 0.04 0.0001 0.001 0.001

Log2 Fold Change 0.005

(Gene / 0.0005 0.01 0.02 b 0 0 0 0 0 0 0 WGE p35S::vector p35S::GmNAC42-1

GmNAC42-1 IFS1 IFS2 I2'H G4DT a GmNAC42-2 GmNAC42-3 0.06 0.07 0.1 0.08 0.012

) a a 0.05 0.0007 a 0.06 0.07 0.05 0.08 0.01 0.0006 0.05 0.04 0.06 0.04 0.008 0.0005 0.06 0.05 0.04 0.03 0.0004 0.03 0.04 0.006

UBIQUIITIN3 0.03 0.04 0.03 0.0003 0.02 0.02 0.004 0.02 0.02 0.0002 0.02 0.01 Log2 Fold Change 0.01 0.01 0.01 0.002 0.0001 (Gene / c 0 0 0 0 0 0 0 H2O p35S::vector p35S::GmNAC42-1 GmNAC42-1 IFS1 IFS2 I2'H G4DT GmNAC42-2 GmNAC42-3 ) 0.02 0.012 0.009 0.0014 0.014 0.00003 0.008 0.01 0.0012 0.012 0.0003 0.016 0.007 0.01 0.008 0.006 0.001 0.00002 0.012 0.005 0.0008 0.008 0.0002

UBIQUIITIN3 0.006 0.008 0.004 0.0006 0.006 0.004 0.003 0.0004 0.004 0.0001 0.00001 Log2 Fold Change 0.004 0.002

(Gene / 0.002 0.001 0.0002 0.002 0 0 0 0 0 0 0 d a Inset 140 25 a FW)

-1 20 RNAi-vector, H2O 120 15 RNAi-GmNAC42-1, H2O 10 b a aa RNAi-vector, WGE 100 5 b b RNAi-GmNAC42-1, WGE aaaa a a b FW) 0 cc cc b aaa 1 - 80 Amount (µg gt a a c 60 a c 40 Amount (µg gt b a a a b b 20 b b a c ab a b bb aaaa 0 c

e a a p35S::vector, H2O 140 a Inset 8

FW) p35S::GmNAC42-1, H2O 1 - 120 6 p35S::vector, WGE 4 b p35S::GmNAC42-1, WGE a 100 a 2 b b cc aaaa aaaa aaaa FW) 0 cc a -1 80 Amount (µg gt b b b c 60 b a a 40 c ba Amount (µg gt b a a 20 b a a a b a a b b b b c b b a 0 b

Fig. 6 (See legend on next page.) Jahan et al. BMC Genomics (2019) 20:149 Page 16 of 21

(See figure on previous page.) Fig. 6 Overexpression and silencing of GmNAC42–1 in soybean hairy roots. a Gene expressions in WGE-treated Williams 82 hairy roots undergoing RNAi silencing of GmNAC42–1. b Gene expressions in WGE-treated hairy roots overexpressing GmNAC42–1. c Gene expressions in mock-treated hairy roots overexpressing GmNAC42–1. Measurements were 24 h after treatment by qRT-PCR. aSignificantly greater and bsignificantly less than control, paired students t-test (P <0.01).d Amounts of phytoalexins and constitutive isoflavonoids in soybean hairy roots undergoing RNAi silencing of GmNAC42–1 24 h after treatment with WGE or H2O. e Metabolite amounts from hairy roots overexpressing GmNAC42–1. Different letters show significant differences by single factor ANOVA, Tukey post hoc test, P < 0.01

N-terminal GFP tag and expressed the translational fusion The nGFP-GmNAC42–1 fusion protein localized to the in soybean hairy roots using the constitutively active nucleus in the absence of an elicitor treatment and thus CaMV-35S promoter (p35S) [53]. nGFP-GmNAC42–1lo- did not rely on elicitor treatment for nuclear localization calized to the nucleus as shown by co-localization with pro- as observed for the phytoalexin TF AtWRKY33 or the pidium iodide fluorescence (red arrowheads, Fig. 7a–c). By NAC-family TFs StNTP1 and StNTP2 [55, 56]. Since contrast, GFP expressed by the empty vector localized to GmNAC42–1 is essential for full elicitation of glyceollins, the cytosol and other extra-nuclear compartments we suggest that GmNAC42–1 acts in concert with at least (Fig. 7d–f). one other TF to coordinately activate all glyceollin biosyn- To test whether the GmNAC42–1 protein could dir- thetic genes. Further, by upregulating some but not all gly- ectly bind the promoters of glyceollin biosynthesis genes, ceollin genes, GmNAC42–1 could also function in SAR to the ORF was also cloned downstream of the GAL4 acti- prime soybean tissues distal to an inoculation site for sub- vation domain and expressed in yeast harboring several sequent rapid/high-level elicitation [23, 57, 58]. A subse- 500 bp segments of IFS2 or G4DT promoters (Fig. 7g). quent direct inoculation of the primed tissues would GmNAC42–1 weakly activated the G4DT promoter seg- activate the expressions or activity of one or more add- ment closest the transcription start site (G4DTpro1)that itional TFs that upregulates I2’H and other glyceollin bio- had one predicted NAC binding element (T/ATTGACT/ synthesis genes that are not regulated by GmNAC42–1 C), failed to activate the segment that lacked the element alone. In that case, overexpressing GmNAC42–1 could (G4DTpro1), and strongly activated both IFS2 promoter serve as an alternative to spraying the lactofen-containing segments that each had several elements (Fig. 7h). herbicide Cobra that primes glyceollin biosynthesis to in- crease resistance against pathogens such as white mold, Discussion the causal agent of sclerotinia stem rot, without adversely GmNAC42–1 is required for full elicitation of glyceollin effecting yield [59, 60]. Future experiments should test biosynthesis whether overexpressing GmNAC42–1 in soybean plants In this study, we found that transcripts of the NAC-family primes glyceollin biosynthesis without adverse effects on TF gene GmNAC42–1 were upregulated with glyceollin yield as well. Since the rapidity of glyceollin elicitation is a biosynthesis genes and metabolites when soybean tissues major factor that distinguishes resistant to P. sojae (Rps) were elicited by acidity stress or the biotic elicitor WGE soybean genotypes from nearly-isogenic susceptible ge- from P. sojae. They were also downregulated with glyceol- notypes [61–64], experiments should also test whether lin biosynthesis genes and metabolites by dehydration overexpressing GmNAC42–1 enhances the rapidity of stress. The overexpression and silencing of GmNAC42–1 glyceollin elicitation in response to compatible P. sojae in WGE-treated hairy roots enhanced and suppressed, re- (Rps) genotypes. spectively, the expressions of the isoflavone biosynthetic genes IFS1 and IFS2, the glyceollin-specific gene G4DT, GmNAC42–1 and a conserved phytoalexin elicitation and the accumulation of glyceollin metabolites. Since pathway G4DT is specifically involved in glyceollin biosynthesis, The regulation of phytoalexins by pathogens and specific the results suggest that GmNAC42–1 is a regulator of gly- abiotic stresses suggests that elicitation is highly complex ceollin elicitation and not the biosynthesis of constitutively and may require multiple signaling pathways. This study accumulating isoflavone conjugates. However, overex- in soybean identified acidity stress (pH 3.0 medium) and pressing or silencing GmNAC42–1 did not affect the dehydration as novel regulators of phytoalexin biosyn- expression levels of I2’H, one of the key genes required for thesis. Transcriptome analysis found that the genes upreg- glyceollin biosynthesis [54]. Further, overexpression of ulated by acidity stress and downregulated by dehydration GmNAC42–1 in the absence of WGE did not result in the were reminiscent of pathogen responses, with SAR genes accumulation of glyceollins. Thus, our results showed that being highly overrepresented. The SAR genes included GmNAC42–1 is required for the full elicitation of glyceol- homologs of Arabidopsis ALD1 and FMO1 that synthesize linbiosynthesisinresponsetoP. sojae WGE, but is not the systemic signaling molecules Pip and its derivative sufficient to upregulate all glyceollin biosynthesis genes. N-hydroxypipecolic acid (NHP) to orchestrate priming of Jahan et al. BMC Genomics (2019) 20:149 Page 17 of 21

a d

b e

c f

g G4DTpro2 G4DTpro1 N Glyma.10G295300

IFS2pro2 IFS2pro1 N N N N N N 125 bp -888 -715 -402 -382 -209 -153 Glyma.13G173500 h Bait 3-AT (mM) Vector GmNAC42-1 G4DTpro1::HIS3 0

40

G4DTpro2::HIS3 0

20

IFS2pro1::HIS3 0 20

IFS2pro2::HIS3 0 10

Fig. 7 Nuclear localization and DNA binding activities of GmNAC42–1. (a-f) Confocal fluorescence microscopy images of GFP-GmNAC42–1 fusion protein in transgenic soybean hairy roots. a-c Root cell expressing GFP-GmNAC42–1 fusion protein. d-f Root cell expressing GFP. a Propidium iodide (10 μgmL− 1) staining the plasma membrane and nucleus. b GFP-GmNAC42–1 fluorescence at plasma membrane and nucleus. c Overlay of GmNAC42–1 and propidium iodide fluorescence from panels a and b. d Propidium iodide staining. e GFP signal in cytosol and other extranuclear compartments. f Overlay of GFP and propidium iodide fluorescence from D and E. Bars = 10 μm. g Schematic diagram demonstrating G4DT and IFS2 promoter fragments used for yeast one-hybrid assays and predicted NAC binding elements (blue boxes). h Yeast one-hybrid assays of YM4271 yeast transformed with GmNAC42–1 on SD/−His/−Leu medium containing various concentrations of 3-AT. Arrows: gray indicates weak binding, white no binding, and black strong binding pathogen responses [46, 47], and the lipase-like and TF [7] were upregulated by long-term acidity stress, sug- Arabidopsis genes PAD4 and ANAC042 that regulate the gesting that NAC42-dependent induction of phyto- biosynthesis of camalexin in Arabidopsis [7, 48]. Here, we alexins may be a conserved response to acidity stress. found that GmNAC42–1 is the soybean homolog of More insight into the NAC42 pathway could be drawn ANAC042 and is required for full elicitation of glyceollins. from the fact that glyceollin biosynthesis was elicited by The results suggest a conserved phytoalexin elicitation the treatment of soybean cotyledons with hydroxyl rad- pathway for phenylpropanoid-derived glyceollins in soy- ical (a ROS) [24] and camalexin elicitation by the bean and indole alkaloid-derived camalexin in Arabidopsis ROS-inducing herbicide acifluorofen required ANAC042 that requires NAC42 TFs. Further, our investigation of [7]. The ROS-inducing herbicide lactofen systemically Lager’stranscriptomedataset[65]demonstratedthat primes glyceollin biosynthesis [59]. ROS accumulation is ANAC042 and its target camalexin biosynthesis genes stimulated by various phytoalexin elicitors such as path- (namely CYP71A12, CYP71A13 and CYP71B15/PAD3) ogens, heavy metals, and UV irradiation [66–68]. Jahan et al. BMC Genomics (2019) 20:149 Page 18 of 21

Further, the acidification of growth media from pH 5.0 experiments demonstrated that GmNAC42–1 regulated to 4.5 stimulated ROS production in seedlings of barley isoflavonoid- and glyceollin-specific biosynthetic genes and Scots pine [69, 70] and in MS medium containing through the direct binding of their promoters. While Plantago shoots [71]. Also, genes that positively regulate our promoter sequence analyses identified the putative ROS (GO:2000377 and GO:2000379) were overrepre- NAC-binding element T/ATTGACT/C within 1 kb of sented in the soybean and Arabidopsis transcriptome re- the translation start sites of the camalexin-specific bio- sponses to long-term acidity stress. Thus, the NAC42 synthetic genes CYP71A12 and CYP71A13 that were pathway may be a conserved ROS signaling pathway regulated at the mRNA level by ANAC042 [7], the responsible for phytoalexin elicitation in response to vari- DREB2A element that was suggested to be the target of ous abiotic and biotic elicitors. It is tempting to speculate ANAC042/JUB1 [76] was not found in those regions that major TFs that regulate acidity and dehydration re- nor within glyceollin biosynthetic gene promoters. If sponses may regulate GmNAC42–1 since the stresses op- NAC42 TFs indeed bind the element T/ATTGACT/C positely regulate GmNAC42–1 transcripts. STOP1 is a element in glyceollin- and camalexin-specific biosyn- zinc finger TF that is a major regulator of protective re- thetic genes, this would suggest that phytoalexin bio- sponses to acidity stress [72, 73]. STOP1 also stimulates synthesis pathways were co-opted into stress-inducible ROS production [74]. Yet, STOP1 homologs were not regulation by NAC42 TFs. Our future work will focus found in the soybean transcriptome response to long-term on characterizing the recognition elements and DNA acidity stress (9 dat), and ANAC042 was not downregu- binding domains of GmNAC42–1 and ANAC042 that lated in an Arabidopsis stop1 mutant at 1 dat [72]. This are required to activate phytoalexin biosynthesis. could infer that NAC42 induction of phytoalexins is downstream of ROS signaling and not directly regulated Conclusions by STOP1. ABA is a major regulator of dehydration re- GmNAC42–1 is essential for the full elicitation of glyceol- sponses in part through the activity of ABA-responsive lins in soybean. It’soverexpressioninelicitedsoybean element (ABRE)-binding TFs [75]. Our transcriptome hairy roots enhanced the biosynthesis of glyceollins more dataset shows that dehydration is a powerful negative than 10-fold. Thus, bioengineering the expressions of regulator of glyceollin biosynthesis and GmNAC42–1, GmNAC42–1 may be a promising approach for bioprodu- raising the possibility that both are negatively regulated by cing glyceollins for medicinal use or for enhancing soy- ABA. We found that ABREs were present in the promoter bean resistance to the economically destructive pathogen regions (~ 1000 bp upstream of the transcription start site) P. sojae.GmNAC42–1 is the first identified conserved of several glyceollin biosynthesis genes, but no ABREs regulator of phytoalexin biosynthesis and is a homolog of were observed in the GmNAC42–1 promoter (data not the indole alkaloid phytoalexin regulator ANAC042 from shown). Thus, dehydration may regulate glyceollin biosyn- Arabidopsis. Possible implications are that NAC42-type thesis at multiple levels. TF genes could be used in a wide variety of crop plants to enhance the bioproduction of medicinal metabolites or for Co-option of phytoalexin biosynthesis by NAC42 improving crop resistance to pathogens. Phytoalexin TF genes of the NAC, MYB, bHLH, and WRKY families have been identified from Arabidopsis, rice, Additional files cotton, maize and grapevine [5–10]. Yet none of these TF genes were homologous among plant species. The phyto- Additional file 1: Table S1. Sequences of primers used in the alexins elicited in these species were biosynthetically diverse experiments. (XLSX 11 kb) and included indole alkaloids, momilactones and phytocas- Additional file 2: Table S2. Sequences of promoters used in yeast sanes, terpenoid aldehydes, deoxyanthocyanidins, and stil- one-hybrid experiments. (XLSX 12 kb) benoids, respectively. Thus, it has remained a question Additional file 3: Table S3. Genes upregulated by pH 3.0 medium compared to control at 9 dat in Harosoy 63 soybean seedlings. whether any phytoalexin TFs are conserved in plants or (XLSX 193 kb) whether they are as diverse as the biosynthetic pathways Additional file 4: Table S4. Genes downregulated by dehydration in that they regulate. Here, we found that GmNAC42–1 is re- Harosoy 63 soybean seedlings compared to the control at 6 days after quired for the full activation of glyceollin biosynthesis in treatment. (XLSX 503 kb) soybean. Its homolog ANAC042 is needed for the full elicit- Additional file 5: Table S5. Genes upregulated by pH 3.0 medium compared to control at 9 dat and downregulated by dehydration compared ation of camalexin biosynthesis in Arabidopsis [7]. Glyceol- to control at 6 dat in Harosoy 63 soybean seedlings. (XLSX 105 kb) lins are isoflavonoid derivatives derived from phenylalanine, Additional file 6: Table S6. Amino acid similarities of NAC42 proteins whereas camalexin is an indole alkaloid biosynthesized from soybean, Arabidopsis and grapevine. Full-length proteins (N-terminal from tryptophan. It is possible that NAC42 TFs regulate and C-terminal halves). (XLSX 11 kb) genes in the shikimate pathway that produces phenylalan- Additional file 7: Figure S1. Amino acid alignment of NAC42 proteins from soybean, Arabidopsis and grapevine. (DOCX 14 kb) ine and tryptophan. Yet, our overexpression and silencing Jahan et al. BMC Genomics (2019) 20:149 Page 19 of 21

Abbreviations Author details ABRC: Arabidopsis Biological Resource Center; ABRE: ABA-responsive element; 1Division of Plant and Soil Sciences, West Virginia University, Morgantown, AgNO3: silver nitrate; ALD1: AGD2-LIKE DEFENSE RESPONSE PROTEIN 1; West Virginia 26506, USA. 2Department of Biology, West Virginia University, aos: allene oxide cyclase; CAD1: (+)-δ-cadinene synthase; CUC2: Cup-shaped Morgantown, West Virginia 26506, USA. 3Department of Biochemistry, West cotyledon; CuCl2: Copper chloride; DMAPP: Dimethylallyl diphosphate; Virginia University, Morgantown, West Virginia 26506, USA. 4Department of DW: Dry weight; FDR: False discovery rate; FMO1: FLAVIN-DEPENDENT- Biostatistics, West Virginia University, Morgantown, West Virginia 26506, USA. MONOOXYGENASE1; FW: Fresh weight; G4DT: Glycinol:4-dimethylallyl 5Microscope Imaging Facility, West Virginia University, Morgantown, West diphosphate transferase; G4DTpro1: G4DT gene promoter segment1; Virginia 26506, USA. GC: Germination and co-cultivation; GRN: Gene regulatory network; HRG: Hairy root growth; MeJA: Methyl jasmonate; MEP: Methylerythritol Received: 24 September 2018 Accepted: 11 February 2019 phosphate; MS: Murashige and Skoog; NAM: No apical meristem; nGFP: N- terminal GFP tag; NHP: N-hydroxypipecolic acid; PAD4: PHYTOALEXIN DEFICIENT4; PAL: Phenylalanine ammonia lyase; PaNie: Pythium aphanidermatum; Pip: Pipecolic acid; qRT-PCR: quantitative reverse References transcriptase-polymerase chain reaction; RIN: RNA Integrity Number; 1. Müller KO, Meyer G, Klinkowski M. Physiologisch-genetische RNAi: RNA interference; RNA-seq: RNA sequencing; ROS: Reactive oxygen Untersuchungen über die Resistenz der Kartoffel gegenüber Phytophthora – species; Rps: Resistant to P. sojae; SAR: Systemic acquired resistance; SD- infestans. Naturwissenschaften. 1939;27(46):765 8. His: Synthetic dropout medium lacking histidine; SIB1: SIGMA FACTOR 2. Graham TL, Graham MY, Subramanian S, Yu O. RNAi silencing of genes for BINDING PROTEIN 1; STS: STILBENE SYNTHASE; TF: Transcription factor; elicitation or biosynthesis of 5-deoxyisoflavonoids suppresses race-specific WGE: Wall glucan elicitor; Y1H: Yeast one-hybrid; 3-AT: 3-amino-1,2,4-triazole; resistance and hypersensitive cell death in Phytophthora sojae infected – ATAF1/2: Arabidopsis transcription activation factor; NAC: NAM/ATAF1/2/ tissues. Plant Physiol. 2007;144(2):728 40. CUC2; osjar1–2: Oryza sativa jasmonic acid-amido synthetase 3. Lygin AV, Zernova OV, Hill CB, Kholina NA, Widholm JM, Hartman GL, Lozovaya VV. Glyceollin is an important component of soybean plant defense against Phytophthora sojae and Macrophomina phaseolina. Acknowledgements Phytopathology. 2013;103(10):984–94. We would like to acknowledge the WVU Genomics Core Facility, Morgantown 4. Yoshikawa M, Yamauchi K, Masago H. Glyceollin: its role in restricting fungal WV for support provided to help make this publication possible. We thank growth in resistant soybean hypocotyls infected with Phytophthora Gustavo MacIntosh and Jessica Hohenstein (Iowa State University) for megasperma var. sojae. Physiol Plant Pathol. 1978;12(1):73–82. Agrobacterium rhizogenes strain K599, Wayne Parrott and Tim Chappell 5. Ibraheem F, Gaffoor I, Tan Q, Shyu C-R, Chopra S. A sorghum MYB (University of Georgia) for the soybean hairy root transformation protocol, Brett transcription factor induces 3-deoxyanthocyanidins and enhances resistance P. sojae Tyler (Oregon State University) for race 1 , Erich Grotewold (University of against leaf blights in maize. Molecules. 2015;20(2):2388–404. Michigan) for the yeast YM4271, Tsuyoshi Nakagawa (Shimane University) for 6. Ogawa S, Miyamoto K, Nemoto K, Sawasaki T, Yamane H, Nojiri H, Okada K. the pGWB2 vector, the ABRC (Columbus, OH) for pDEST-GADT7, and Hiroyuki OsMYC2, an essential factor for JA-inductive sakuranetin production in rice, Tsuji (Yokohama City University) and the late Ko Shimamoto for pANDA35HK. interacts with MYC2-like proteins that enhance its transactivation ability. Sci n We acknowledge use of the WVU Shared Research Facilities UPLC-PDA-MS . Rep. 2017;7:40175. Imaging experiments were performed at the West Virginia University 7. Saga H, Ogawa T, Kai K, Suzuki H, Ogata Y, Sakurai N, Shibata D, Ohta D. Microscope Imaging Facility, which has been supported by the WVU Identification and characterization of ANAC042, a transcription factor family Cancer Institute and NIH grants P20RR016440 and P30RR032138, gene involved in the regulation of camalexin biosynthesis in Arabidopsis. P30GM103488 and U54GM104942 for the Nikon A1R/N SIM-E. Mol Plant Microbe In. 2012;25(5):684–96. 8. Xu Y-H, Wang J-W, Wang S, Wang J-Y, Chen X-Y. Characterization of Funding GaWRKY1, a cotton transcription factor that regulates the sesquiterpene This work was supported by WVU start-up funds to NK and was based upon synthase gene (+)-δ-cadinene synthase-a. Plant Physiol. 2004;135(1):507–15. work that is supported by the USDA National Institute of Food and Agricul- 9. Yamamura C, Mizutani E, Okada K, Nakagawa H, Fukushima S, Tanaka A, ture, Hatch, accession number 1010200, and the West Virginia Agricultural Maeda S, Kamakura T, Yamane H, Takatsuji H, et al. Diterpenoid Phytoalexin and Forestry Experiment Station project WVA00687. factor, a bHLH transcription factor, plays a central role in the biosynthesis of Diterpenoid Phytoalexins in Rice. Plant J. 2015;84(6):1100–13. 10. Zheng Z, Qamar SA, Chen Z, Mengiste T. Arabidopsis WRKY33 transcription Availability of data and materials factor is required for resistance to necrotrophic fungal pathogens. Plant J. All RNA-seq data are available in the Gene Expression Omnibus (https:// 2006;48(4):592–605. www.ncbi.nlm.nih.gov/geo/) under the series accession GSE112584. 11. Ahuja I, Kissen R, Bones AM. Phytoalexins in defense against pathogens. Trends Plant Sci. 2012;17(2):73–90. Authors’ contributions 12. Großkinsky DK, van der Graaff E, Roitsch T. Phytoalexin transgenics in crop n M.A.J. conducted UPLC-PDA-MS and statistics, B.H., K.C., and M.A.J. soybean protection—fairy tale with a happy end? Plant Sci. 2012;195:54–70. hairy root transformations, M.L. qRT-PCR, B.H. Y1H; A.G.A. microscopy, R.J.P. 13. Jeandet P: Phytoalexins: current progress and future prospects. In: RNA-seq library preparation and A.M.I. processed data. N.K. cloning, yeast Multidisciplinary Digital Publishing Institute; 2015. transformations, supervised research, analyzed data, and wrote the manu- 14. Marti G, Schnee S, Andrey Y, Simoes-Pires C, Carrupt P-A, Wolfender J-L, Gindro script. All authors read and approved the final manuscript. K. Study of leaf metabolome modifications induced by UV-C radiations in representative Vitis, Cissus and Cannabis species by LC-MS based Ethics approval and consent to participate metabolomics and antioxidant assays. Molecules. 2014;19(9):14004–21. Not applicable. 15. Park HL, Lee S-W, Jung K-H, Hahn T-R, Cho M-H. Transcriptomic analysis of UV- treated rice leaves reveals UV-induced phytoalexin biosynthetic pathways and their regulatory networks in rice. Phytochemistry. 2013;96:57–71. Consent for publication 16. Kato H, Kodama O, Akatsuka T. Oryzalexin E, a diterpene phytoalexin from No applicable. UV-irradiated rice leaves. Phytochemistry. 1993;33(1):79–81. 17. Mert-Turk F, Bennett MH, Mansfield JW, Holub EB. Quantification of Competing interests camalexin in several accessions ofArabidopsis thaliana following The authors declare that they have no competing interests. inductions withPeronospora parasitica and UV-B irradiation. Phytoparasitica. 2003;31(1):81–9. 18. Reilly JJ, Klarman WL. Thymine dimer and glyceollin accumulation in UV- Publisher’sNote irradiated soybean suspension cultures. Environ Exp Bot. 1980;20(2):131–4. Springer Nature remains neutral with regard to jurisdictional claims in 19. Miyamoto K, Enda I, Okada T, Sato Y, Watanabe K, Sakazawa T, Yumoto E, published maps and institutional affiliations. Shibata K, Asahina M, Iino M. Jasmonoyl-L-isoleucine is required for the Jahan et al. BMC Genomics (2019) 20:149 Page 20 of 21

production of a flavonoid phytoalexin but not diterpenoid phytoalexins in 43. Mishina TE, Zeier J. The Arabidopsis flavin-dependent monooxygenase ultraviolet-irradiated rice leaves. Biosci Biotechnol Biochem. 2016;80(10):1934–8. FMO1 is an essential component of biologically induced systemic acquired 20. Rakwal R, Tamogami S, Kodama O. Role of jasmonic acid as a signaling resistance. Plant Physiol. 2006;141(4):1666–75. molecule in copper chloride-elicited rice phytoalexin production. Biosci 44. Návarová H, Bernsdorff F, Döring A-C, Zeier J. Pipecolic acid, an endogenous Biotechnol Biochem. 1996;60(6):1046–8. mediator of defense amplification and priming, is a critical regulator of 21. Farrell KC, Jahan MA, Kovinich N. Distinct mechanisms of biotic and inducible plant immunity. Plant Cell. 2012;24(12):5123–41. chemical elicitors enable additive elicitation of the anticancer Phytoalexin 45. Song JT, Lu H, Greenberg JT. Divergent roles in Arabidopsis thaliana Glyceollin I. Molecules. 2017;22(8):1261–73. development and defense of two homologous genes, aberrant growth and 22. Beyer EM. A potent inhibitor of ethylene action in plants. Plant Physiol. death2 and AGD2-LIKE DEFENSE RESPONSE PROTEIN1, encoding novel 1976;58(3):268–71. aminotransferases. Plant Cell. 2004;16(2):353–66. 23. Park D, Landini S, Graham M, Graham T. Induced distal defence potentiation 46. Hartmann M, Zeier T, Bernsdorff F, Reichel-Deland V, Kim D, Hohmann M, against Phytophthora sojae in soybean. Physiol Mol Plant Pathol. 2002;60(6): Scholten N, Schuck S, Bräutigam A, Hölzel T. Flavin monooxygenase- 293–310. generated N-hydroxypipecolic acid is a critical element of plant systemic 24. Epperlein M, Noronha-Dutra A, Strange R. Involvement of the hydroxyl immunity. Cell. 2018;173(2):456–69 e416. radical in the abiotic elicitation of phytoalexins in legumes. Physiol Mol 47. Bernsdorff F, Döring A-C, Gruner K, Schuck S, Bräutigam A, Zeier J. Pipecolic Plant Pathol. 1986;28(1):67–77. acid orchestrates plant systemic acquired resistance and defense priming 25. Qiu JL, Fiil BK, Petersen K, Nielsen HB, Botanga CJ, Thorgrimsen S, Palma K, via salicylic acid-dependent and-independent pathways. Plant Cell. Suarez-Rodriguez MC, Sandbech-Clausen S, Lichota J. Arabidopsis MAP 2016;28(1):102–29. kinase 4 regulates gene expression through transcription factor release in 48. Ferrari S, Plotnikova JM, De Lorenzo G, Ausubel FM. Arabidopsis local the nucleus. EMBO J. 2008;27(16):2214–21. resistance to Botrytis cinerea involves salicylic acid and camalexin and 26. Höll J, Vannozzi A, Czemmel S, D'Onofrio C, Walker AR, Rausch T, Lucchin M, requires EDS4 and PAD2, but not SID2, EDS5 or PAD4. Plant J. 2003;35(2): Boss PK, Dry IB, Bogs J. The R2R3-MYB transcription factors MYB14 and MYB15 193–205. regulate stilbene biosynthesis in Vitis vinifera. Plant Cell. 2013;25(10):4135–49. 49. Lai Z, Li Y, Wang F, Cheng Y, Fan B, Yu J-Q, Chen Z. Arabidopsis sigma 27. Chen Y, Chen Z, Kang J, Kang D, Gu H, Qin G. AtMYB14 regulates cold factor binding proteins are activators of the WRKY33 transcription factor in tolerance in Arabidopsis. Plant Mol Biol Report. 2013;31(1):87–97. plant defense. Plant Cell. 2011;23(10):3824–41. 28. Chezem WR, Memon A, Li F-S, Weng J-K, Clay NK. SG2-type R2R3-MYB 50. Toth Z, Winterhagen P, Kalapos B, Su Y, Kovacs L, Kiss E. Expression of transcription factor MYB15 controls defense-induced lignification and basal a grapevine NAC transcription factor gene is induced in response to immunity in Arabidopsis. Plant Cell. 2017;29(8):1907–26. powdery mildew colonization in salicylic acid-independent manner. Sci 29. Frerigmann H, Glawischnig E, Gigolashvili T. The role of MYB34, MYB51 and Rep. 2016;6:30825. MYB122 in the regulation of camalexin biosynthesis in Arabidopsis thaliana. 51. Thirumalaikumar VP, Devkar V, Mehterov N, Ali S, Ozgur R, Turkan I, Mueller- Front Plant Sci. 2015;6:654. Roeber B, Balazadeh S. NAC transcription factor JUNGBRUNNEN1 enhances 30. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, drought tolerance in tomato. Plant Biotechnol J. 2017. Chaisson M, Gingeras TR. STAR: ultrafast universal RNA-seq aligner. 52. Yang Y, Zhu K, Wu J, Liu L, Sun G, He Y, Chen F, Yu D. identification and Bioinformatics. 2013;29(1):15–21. characterization of a novel NAC. Plant Cell Rep. 2016;35(8):1783–98. 31. Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose 53. Odell JT, Nagy F, Chua N-H. Identification of DNA sequences required for program for assigning sequence reads to genomic features. Bioinformatics. activity of the cauliflower mosaic virus 35S promoter. Nature. 1985; 2013;30(7):923–30. 313(6005):810. 32. Anders S, Huber W. Differential expression analysis for sequence count data. 54. Akashi T, Aoki T, Ayabe S-i. CYP81E1, a cytochrome P450 cDNA of licorice Genome Biol. 2010;11(10):R106. (Glycyrrhiza echinataL.), encodes Isoflavone 2′-hydroxylase. Biochem Biophys 33. Jacobs TB, LaFayette PR, Schmitz RJ, Parrott WA. Targeted genome Res Commun. 1998;251(1):67–70. modifications in soybean with CRISPR/Cas9. BMC Biotechnol. 2015;15(1):16. 55. Mao G, Meng X, Liu Y, Zheng Z, Chen Z, Zhang S. Phosphorylation of a 34. Yang F, Ouma W, Li W, Doseff A, Grotewold E: Establishing the architecture WRKY transcription factor by two pathogen-responsive MAPKs drives of plant gene regulatory networks. In: Methods Enzymol vol. 576: Elsevier; phytoalexin biosynthesis in Arabidopsis. Plant Cell. 2011;23(4):1639–53. 2016: 251–304. 56. McLellan H, Boevink PC, Armstrong MR, Pritchard L, Gomez S, Morales J, 35. Bhattacharyya M, Ward E. Resistance, susceptibility and accumulation of Whisson SC, Beynon JL, Birch PR. An RxLR effector from Phytophthora glyceollins I–III in soybean organs inoculated with Phytophthora infestans prevents re-localisation of two plant NAC transcription factors megasperma f. Sp. glycinea. Physiol Mol Plant Pathol. 1986;29(2):227–37. from the endoplasmic reticulum to the nucleus. PLoS Pathog. 2013;9(10): 36. Akashi T, Aoki T, Ayabe S-i. Molecular and biochemical characterization of 2- e1003670. hydroxyisoflavanone dehydratase. Involvement of carboxylesterase-like proteins 57. Graham T, Graham M. Glyceollin elicitors induce major but distinctly in leguminous isoflavone biosynthesis. Plant Physiol. 2005;137(3):882–91. different shifts in isoflavonoid metabolism in proximal and distal soybean 37. Akashi T, Sasaki K, Aoki T, Ayabe S-i, Yazaki K. Molecular cloning and cell populations. Mol Plant-Microbe Interact. 1991;4:60–8. characterization of a cDNA for pterocarpan 4-dimethylallyltransferase 58. Subramanian S, Graham MY, Yu O, Graham TL. RNA interference of soybean catalyzing the key prenylation step in the biosynthesis of glyceollin, a isoflavone synthase genes leads to silencing in tissues distal to the soybean phytoalexin. Plant physiology. 2009;149(2):683–93. transformation site and to enhanced susceptibility to Phytophthora sojae. 38. Yoneyama K, Akashi T, Aoki T. Molecular characterization of soybean Plant Physiol. 2005;137(4):1345–53. Pterocarpan 2-Dimethylallyltransferase in Glyceollin biosynthesis: local gene 59. Landini S, Graham MY, Graham TL. Lactofen induces isoflavone accumulation and Whole-genome duplications of Prenyltransferase genes led to the and glyceollin elicitation competency in soybean. Phytochemistry. 2003;62(6): structural diversity of soybean Prenylated Isoflavonoids. Plant Cell Physiol. 865–74. 2016;57(12):2497–509. 60. Dann E, Diers B, Hammerschmidt R. Suppression of Sclerotinia stem rot of 39. Chu S, Wang J, Zhu Y, Liu S, Zhou X, Zhang H, C-e W, Yang W, Tian Z, soybean by lactofen herbicide treatment. Phytopathology. 1999;89(7):598–602. Cheng H. An R2R3-type MYB transcription factor, GmMYB29, regulates 61. Graham T, Kim J, Graham M. Role of constitutive isoflavone conjugates in isoflavone biosynthesis in soybean. PLoS Genet. 2017;13(5):e1006770. the accumulation of glyceollin in soybean infected with Phytophthora 40. Zhao M, Wang T, Wu P, Guo W, Su L, Wang Y, Liu Y, Yan F, Wang Q. Isolation megasperma. Mol Plant-Microbe Interact. 1990;3:157–66. and characterization of GmMYBJ3, an R2R3-MYB transcription factor that 62. Keen N. Specific elicitors of plant phytoalexin production: detenninants of affects isoflavonoids biosynthesis in soybean. PLoS One. 2017;12(6):e0179990. race specificity in pathogens? Science. 1975;187(4171):74–5. 41. Yi J, Derynck MR, Li X, Telmer P, Marsolais F, Dhaubhadel S. A single-repeat 63. Klarman WL, Gerdemann JW. Resistance of to three Phytophthora MYB transcription factor, GmMYB176, regulates CHS8 gene expression and species due to the production of a phytoalexin. Phytopathology. 1963;53: affects isoflavonoid biosynthesis in soybean. Plant J. 2010;62(6):1019–34. 1317–20. 42. Mainali HR, Vadivel AKA, Li X, Gijzen M, Dhaubhadel S. Soybean cyclophilin 64. Yoshikawa M, Masago H: Biochemical mechanism of glyceollin GmCYP1 interacts with an isoflavonoid regulator GmMYB176. Sci Rep. accumulation in soybean. Plant infection: the physiological and biochemical 2017;7:39550. basis/edited by Yasuji Asada[et al] 1982. Jahan et al. BMC Genomics (2019) 20:149 Page 21 of 21

65. Lager I, Andréasson O, Dunbar TL, Andreasson E, Escobar MA, Rasmusson AG. Changes in external pH rapidly alter plant gene expression and modulate auxin and elicitor responses. Plant Cell Environ. 2010;33(9):1513–28. 66. Ma Z, Song T, Zhu L, Ye W, Wang Y, Shao Y, Dong S, Zhang Z, Dou D, Zheng X. A Phytophthora sojae glycoside hydrolase 12 protein is a major virulence factor during soybean infection and is recognized as a PAMP. Plant Cell. 2015;27(7):2057–72. 67. Vishwakarma K, Upadhyay N, Singh J, Liu S, Singh VP, Prasad SM, Chauhan DK, Tripathi DK, Sharma S. Differential phytotoxic impact of plant mediated silver nanoparticles (AgNPs) and silver nitrate (AgNO3) on Brassica sp. Front Plant Sci. 2017;8:1501. 68. Mewis I, Schreiner M, Nguyen CN, Krumbein A, Ulrichs C, Lohse M, Zrenner R. UV-B irradiation changes specifically the secondary metabolite profile in broccoli sprouts: induced signaling overlaps with defense response to biotic stressors. Plant Cell Physiol. 2012;53(9):1546–60. 69. Song H, Xu X, Wang H, Tao Y. Protein carbonylation in barley seedling roots caused by aluminum and proton toxicity is suppressed by salicylic acid. Russ J Plant Physiol. 2011;58(4):653–9. 70. Ivanov YV, Savochkin YV, Kuznetsov VV. Effect of mineral composition and medium pH on scots pine tolerance to toxic effect of zinc ions. Russ J Plant Physiol. 2013;60(2):260–9. 71. Martins N, Gonçalves S, Romano A. Metabolism and aluminum accumulation in Plantago almogravensis and P. Algarbiensis in response to low pH and aluminum stress. Biol Plant. 2013;57(2):325–31. 72. Sawaki Y, Iuchi S, Kobayashi Y, Kobayashi Y, Ikka T, Sakurai N, Fujita M, Shinozaki K, Shibata D, Kobayashi M. STOP1 regulates multiple genes that protect Arabidopsis from proton and aluminum toxicities. Plant Physiol. 2009;150(1):281–94. 73. Iuchi S, Koyama H, Iuchi A, Kobayashi Y, Kitabayashi S, Kobayashi Y, Ikka T, Hirayama T, Shinozaki K, Kobayashi M. Zinc finger protein STOP1 is critical for proton tolerance in Arabidopsis and coregulates a key gene in aluminum tolerance. Proc Natl Acad Sci. 2007;104(23):9900–5. 74. Mora-Macías J, Ojeda-Rivera JO, Gutiérrez-Alanís D, Yong-Villalobos L, Oropeza-Aburto A, Raya-González J, Jiménez-Domínguez G, Chávez-Calvillo G, Rellán-Álvarez R, Herrera-Estrella L. Malate-dependent Fe accumulation is a critical checkpoint in the root developmental response to low phosphate. Proc Natl Acad Sci. 2017;114(17):E3563–72. 75. Uno Y, Furihata T, Abe H, Yoshida R, Shinozaki K, Yamaguchi-Shinozaki K. Arabidopsis basic leucine zipper transcription factors involved in an abscisic acid-dependent signal transduction pathway under drought and high- salinity conditions. Proc Natl Acad Sci. 2000;97(21):11632–7. 76. Wu A, Allu AD, Garapati P, Siddiqui H, Dortay H, Zanor M-I, Asensi-Fabado MA, Munné-Bosch S, Antonio C, Tohge T: JUNGBRUNNEN1, a reactive oxygen species–responsive NAC transcription factor, regulates longevity in Arabidopsis. Plant Cell 2012:tpc. 111.090894.